How I know whether my recovered data are good!

[Henry Chua]

Hi,

Can I know whether the data I recovered from the faulty harddisk is not bad?.. If it is down due to damage blocks, how can I verify that the data is not from one of those?..

Thank you for your answers

[Pete Randall]

Have you run fsck on the file systems? If it was successful, then the odds are good that the data is OK.


Pete

[Olivier Decorse · Accepted solution]

Hi Henry,
Usally, under HPUX, when data or link (inode) are bad, they are moved to the lost+found directory of the logical volume.
So if you don't see any file in EVERY lost+found directory (the one of /, the one of /usr, the one of /stand, etc.), then you have better chance to have no corrupted files !!

The only "normal" file you can found in every lost+found directory is a ".fsadm" file, or something like that : is is a lock file for the OnlineJFS product.

Olivier.

[Prashant Zanwar_4]

If it is faulty disk, you can replace it with new disk.

Do a pvcreate a new physical disk- vgxtend to new PV - lvextend to new PV - do a pvmove from old bad disk to new disk. - Do lvreduce of old PV - unmount file system - fsck the FS - mount it back.

You will come to know while fsck if anything is wrong. Any lost file structures will be stored under lost+found. This way you are through to know if you have everything alright. Dont forget to take backup before pvmove.

Hope above helps in determining your query

Thanks
Prashant

[Steven E. Protter]

After verifying the physical integrity of the blocks (fsck and such) you have to test the actual data.

If its an application you need to bring it up and run a global report that covers all the data.

The testing methodology is completely dependent on what was on the disk.

Most people in such situation are going to go back to their last reliable backup and if possible roll the data forward. Unless I have a bulletproof test, I'm not going to rely on the data when corruption could rear its ugly head weeks or months down the road.

SEP

[Bill Hassell]

In general, a faulty disk drive will not allow you backup program to run. If the directory is corrupted, fsck only fixes the directory, not the data. If the backup program runs without an error, both the directory and the data have been read without an error. It is very unusual for a disk failure to allow a backup without errors (most backup programs fail to run if the disk has any errors). There is no way to know if data has been corrupted without any error messages.